Appl Microbiol Biotechnol (1991) 34:559-564 017575989100025C

App//ed Miemb/okgy B/otechnobgy © Springer-Verlag 1991

Chinese hamster ovary cell growth and interferon production kinetics in stirred batch culture P. M. Hayter, E. M. A. Curling, A. J. Baines, N. Jenkins, I. Salmon, P. G. Strange, and A. T. Bull Biological Laboratory, University of Kent, Canterbury, Kent CT2 7NJ, UK Received 21 August 1990/Accepted 5 October 1990

Summary. Recombinant human interferon-7 production by Chinese hamster ovary cells was restricted to the growth phase of batch cultures in serum-free medium. The specific interferon production rate was highest during the initial period of exponential growth but declined subsequently in parallel with specific growth rate. This decline in specific growth rate and interferon productivity was associated with a decline in specific metabolic activity as determined by the rate of glucose uptake and the rates of lactate and ammonia production. The ammonia and lactate concentrations that had accumulated by the end of the batch culture were not inhibitory to growth. Glucose was exhausted by the end of the growth phase but increased glucose concentrations did not improve the cell yield or interferon production kinetics. Analysis of amino acid metabolism showed that glutamine and asparagine were exhausted by the end of the growth phase, but supplementation o f these amino acids did not improve either cell or product yields. When glutamine was omitted from the growth medium there was no cell proliferation but interferon production occurred, suggesting that recombinant protein production can be uncoupled from cell proliferation.

scale culture of mammalian cells for the manufacture of therapeutic proteins. However, product yields in mammalian cell cultures are typically low (approximately 100 mg 1-1 medium consumed) due in part to the relatively low. cell densities and short production phases achieved in mammalian cell culture. Improvements in the economics of mammalian cell production systems are likely, therefore, to be achieved by better medium design and process control. This approach will require a greater understanding of the effect of the environment on cell physiology and recombinant protein production kinetics. Chinese hamster ovary (CHO) cells have been used extensively for the expression of mammalian proteins (Kaufmann et al. 1983, 1988; McCormick et al. 1984) and provide a suitable model for studying the effect of the culture environment on recombinant protein expression. The purpose of this study was to identify factors which influence the production of recombinant human interferon-7~ (IFN-Tz) in a C H O cell line. In particular we wished to determine the effect of the physiological state of C H O cells on both the rate of production and the fidelity of the heterologous protein (Curling et al. 1990). This paper describes preliminary findings on the relationship between growth and interferon production kinetics in C H O cells.

Introduction The production of therapeutic proteins in mammalian cells has advantages over the use of bacteria or yeast due to the more accurate post-translational processing of human proteins in mammalian cells (Bebbington and Hentschel 1985; Lubiniecki 1987). This is particularly important for proteins where correct post-translational modifications are essential for biological activity (Busby et al. 1985; Dube et al. 1988). Consequently, considerable attention has been focused on the large-

Offprint requests to: P. M. Hayter

Materials and methods The cell line used in this investigation was derived from a dihydrofolate reductase deficient (DHFR-) mutant of CHO-K1 cells and was kindly provided by Wellcome Biotechnology (Beckenham, Kent, UK). The cells expressed IFN-~', which was co-amplified with dihydrofolate reductase by methotrexate selection. For this research we have developed a serum-free medium based on Roswell Park Memorial Institute (RPMI) 1640 (Imperial Laboratories, Andover, Hants, UK) and supplemented with 5 mg m1-1 bovine serum albumin (Miles Biochemicals, Slough, Bucks, UK), 5 lxg ml- 1 human transferfin, 5 ktg ml- ~ bovine insulin, 1 mM sodium pyruvate, 0.1 mM alanine, 1 ~M putrescine, 3 gM FeSO4, 3 ~M ZnSO4, 10 nM Na2SeO3 and 10 nM CuSO4 (Sigma, Poole, Dorset, UK). The concentrations of the low-molecularweight nutrients were based on those described by Hamilton and

560 Ham (1977). Methotrexate (Sigma) was added at a final concentration of 0.1 ~tM. CHO cells growing initially in RPMI 1640 ~ontaining 7% adult bovine serum, were selected for growth in serum-free medium. This was achieved over a period of 47 days by a sequence of stepwise reductions in serum concentration in the presence of the supplements described above. The cells were maintained at each stage of selection until the specific growth rate was comparable to that in serum-containing medium. Shake-flask culture studies were made in sealed 250-ml edenmeyer flasks. Inocula for shake-flask cultures were harvested from 500-ml spinner cultures (Techne, Cambridge, UK) and were seeded at 1.5 x 105 cells m1-1 in 100ml serum-free medium. The headspace of the flask was purged with 5% CO2 in air to maintain the pH at 7.2-7.4. Flasks were incubated at 37°C in an orbital shaking incubator at 100 rpm and samples were removed daily for analysis. Fermentor studies were made in a 2-1 fermentor (Bioengineering, Wald, Switzerland) with pH and oxygen control. The dissolved oxygen tension of the medium was controlled at 50% air saturation by sparging with air and the pH was maintained at 7.2 by CO2 addition. The stirrer speed was 60rpm (tip speed, 23.6 cm s-~). Fermentor cultures were inoculated with cells harvested from 500-ml spinner cultures and seeded at 1 × 105 cells m1-1 in 2 1 serum-free medium. Samples were removed twice daily for analysis. All samples for metabolite determinations were stored at -80°C until required. Cell counts were made in a Neubauer counting chamber and cell viability determined by trypan blue dye exclusion. Glucose was determined using the o-toluidine method (Sigma assay no. 635), lactate concentrations by a lactate dehydrogenase procedure (Sigma assay no. 826-UV) and ammonia by the indophenol method of Fawcett and Scott (1960). IFN-~, concentrations were determined by a monoclonal antibody based sandwich enzymelinked immunosorbent assay (ELISA). Assay plates (Falcon, Cockeyville, USA) were coated with 20B8 antibody raised against recombinant IFN-y from Escherichia coli (Celltech, Slough, UK). The plates were washed twice with wash buffer [phosphate-buffered saline (PBS), 0.05% w/v casein, 0.1% v/v Tween 20] and blocked for 1 h in blocking buffer (PBS, 0.5% w/v casein, 0.1% v/v Tween 20). After a further wash, samples and IFN-y standards (Wellcome) were applied diluted in wash buffer. After incubation for 1 h, the plates were washed again and biotinylated 20G7 antibody (Celltech) was added. This antibody, also raised against E. coli-derived IFN-~,, recognises a different epitope on the IFN-y molecule. The plates were incubated for a further hour, washed and then Extravidin-phosphatase (Sigma) was added. After incubating for 1 h the plates were washed for a final time and phosphatase substrate (Sigma 104) was added. The reaction was stopped after 15 min by the addition of 5 M NaOH and the absorbance read at 410 nm in a Dynatech MR5000 plate reader (Dynatech, Billingshurst, West Sussex, UK). Amino acids were derivatised with o-phthalaldehyde (OPA) and analysed by HPLC using a reverse phase C18 column (Hichrum, Reading, Berks., UK). The procedure was similar to that described by Seaver et al. (1984) except that samples for analysis were filtered through 10-kDa exclusion limit ultrafiltration membranes (Millipore, Harrow, Middlesex, UK) and diluted 1:20 in HPLC grade water (Aldrich, Gillingham, Dorset, UK). Samples were derivatised by the addition of 200 I.tl fluoraldehyde (Pierce, Luton, Beds., UK) to 100 ~tl diluted medium. After 2 min the derivatised samples were injected onto the column using a 20-1xlsample loop. The column was calibrated using a standard amino acid mixture (Sigma AA-S-18). Results and discussion C H O cells in fermentors reached maximum viable cell concentrations o f 0.9-1.3 x 106 cells m1-1 after 100 h

(Fig. 1) with a maximum specific growth rate of 0.031 h -~. After 50 h there was a steady decline in the cell growth rate until the end of the growth phase. In these cultures IFN-7/ accumulated during the growth phase only, reaching concentrations o f 4 - 6 x 1 0 3 I U m1-1 (Fig. 1). The maximum specific I F N - ? production rate (qiFr~), which was approximately 200 I U 10 6 cells-1 h - 1 , was associated with the exponential period o f cell growth. After this initial period, the rate of I F N 2, production declined in parallel with the Specific growth rate. Thus there was an apparent relationship between cell growth and IFN-T, production. Glucose was consumed rapidly and was exhausted by 100 h, which coincided with the onset o f the stationary phase. Lactate and ammonia accumulated during growth reaching final concentrations of 12 mM and 2 mM respectively. There was a steady decline in the specific rate o f glucose utilisation (qglucose) and the specific rates o f lactate and ammonia accumulation (qlactate and qa.... ia) during the growth phase. These changes in cell metabolic activity could not be attributed to p H changes or to oxygen starvation since these parameters were maintained at constant levels. A decline in cell growth and metabolic activity may be attributed either to the depletion of an essential nutrient or to the accumulation o f inhibitory metabolites. Two significant end-products o f cell metabolism Were lactic acid and ammonia. Ammonia, which is produced predominantly from glutamine by glutaminase activity (McKeehan 1986), has been reported to inhibit growth o f other cell lines (Holley et al. 1978; Butler and Spier 1984; Glacken et al. 1986; Reuveny et al. 1986). Lactic acid is a product of glucose catab~rlism with a large proportion of glucose being converted to lactate in proliferating cells (Levintow and Eagle 1961 ; Reitzer et al. 1979). Lactic acid may also inhibit cell growth, an effect which has been reported to be independent o f the reduction in culture p H that results from accumulation o f lactic acid (Glacken et al. 1986). The possibility that lactate and a m m o n i a inhibit C H O cell growth was investigated by the addition of varying concentrations of a m m o n i u m chloride or L-lactic acid to 100-ml shake-flask cultures. The p H was corrected by the addition o f N a O H or HC1 to eliminate the possible influence o f p H on growth and metabolism. Growth was not inhibited by initial a m m o n i u m chloride concentrations o f up to 2 m g but was reduced with an initial a m m o n i a concentration o f 4.5 mM (Fig. 2). Furthermore, growth was unaffected by an initial lactate concentration of 17.5 mM, which exceeds the lactate concentrations found in typical batch cultures (Fig. 3). These results suggest that the reduction in specific growth rate in the latter stages o f the growth phase by batch cultures was not due to inhibition by either lactic acid or ammonia. Schlaeger and Schumpp (1989) also f o u n d that C H O cells were relatively insensitive to inhibition by lactate and ammonia. In their study 8-10 mM a m m o n i a and 90-110 mM lactate were required to give 50% inhibition o f C H O cell growth. Alternatively it is possible that the depletion of an essential nutrient was limiting cell growth and, since

561 1'0T

T6

a

15" C

~O-B

-~ ~x ~ 0"6

/

F

~•

Lg ~ ~ ~ 2 ~ =

~

~

~

-0

0

-200C = ~

-150

t:~ ,~_

w ~

-lOO~

~I~-

~- 0"01-50

0

i

i

20

40

i

i

i

BO 100 Time(hours)

60

120

-i U ~

0

1},0 160

2 .~

Chinese hamster ovary cell growth and interferon production kinetics in stirred batch culture.

Recombinant human interferon-gamma production by Chinese hamster ovary cells was restricted to the growth phase of batch cultures in serum-free medium...
590KB Sizes 0 Downloads 0 Views